Concept of PIRO as a new conceptual framework to understand sepsis

Objective: To define and introduce the PIRO model for categorizing sepsis in infants and children.

Design: A summary of the literature published during the 5 yrs since this concept has been formulated, along with a consensus opinion of experts in the field of sepsis and septic shock.

Results: The imprecision and intrinsic heterogeneity of the patient population defined as septic has prompted the introduction of a new sepsis classification system known as PIRO. PIRO stands for predisposition, infection, response, and organ dysfunction. It is hoped that by defining the septic process through a detailed analysis of each of its component parts, the development of sepsis will be better understood. This may contribute to improved therapeutic interventions for sepsis in the future.

Conclusions: The PIRO model is a conceptual framework for understanding sepsis that has many favorable attributes. The PIRO model should be directly tested in both the research laboratory and in clinical trial designs to determine the practical value and clinical relevance of this new classification scheme for sepsis.

Sepsis is a complex, highly variable, multiple system, clinical process induced by microbial pathogens that causes a deleterious, systemic host response. Pathogen-associated molecular patterns such as lipopolysaccharide and peptidoglycan activate innate immune responses, coagulation pathways, and cellular injury and apoptosis gene patterns, culminating in diffuse tissue injury and multiple organ dysfunction. The dynamic nature and complexities of the host-pathogen interactions after systemic infection have proven difficult to categorize into discrete and understandable clinical definitions (¹). The standard working definitions of sepsis, developed in a consensus conference in 1991 (²), have been roundly criticized as being insufficiently specific and reliable to be of clinical value (³).

In recognition of the complexity of sepsis and the inadequacies of current working definitions, a new conceptual framework for analyzing sepsis has been developed, known as PIRO (^{4, 5}). PIRO stands for predisposition, infection, response, and organ dysfunction and was broadly modeled by Dr. John Marshall after the TNM classification for cancer staging (⁴). Oncologists have recognized for decades that cancer classifications are essential not only for prognostic information but also for therapeutic decisions. The TNM classification-tumor size, regional nodal involvement, and local and distant metastases-has greatly facilitated classification and treatment regimens for a variety of human neoplasms and is widely used to direct treatment decisions in the management of cancer. Perhaps a comparable system of nomenclature that recognizes the major component parts of sepsis will improve the understanding and treatment of this common and severe infectious disease (^6-8). It has become abundantly clear that sepsis remains a poorly understood entity that consumes an inordinately large amount of healthcare expenditures in developed countries and continues to account for considerable morbidity and mortality in critical care units (^{5, 9-14}). Despite decades of clinical investigation into improved therapies for sepsis, improvements in outcome have been modest at best, with only one treatment strategy (activated protein C) (¹⁵) succeeding to show benefit in a large, multiple-center, phase III clinical trial (^{8, 12, 16}). The basic elements of the conceptual framework on which the PIRO concept has evolved will be discussed in the following paragraphs. The four major domains of PIRO will be analyzed separately in an attempt to integrate the information into a useful system to analyze sepsis.

P: Predisposing Factors

It is likely that the genetic constitution of an individual is a major contributor to the lifetime risk of severe infection and septic shock. In a long-term follow-up study of >900 Danish families with adopted children, Sorensen et al. (¹⁷) analyzed the risk of death of adoptees in relationship to causes of death in the adoptive parents or the biological parents. If environmental influences primarily determine the risk of mortality from a variety of common illnesses, the adopted children should more likely face the same fate as their adopted parents. If genetic predisposition primarily determines outcome, then the cause of mortality in adopted children should reflect more closely the causes of death in their biological parents. The most striking finding in the study was an almost five-fold increase in the relative risk of death due to infection if either parent had succumbed from a severe infection within the first 50 yrs of life (¹⁷). The associated risk of death related to infection in first-degree relatives was much higher than the associated risk of death from cancer or cardiovascular diseases in their biological parents. No significant association was found between the risk of death due to infection in adopted children and their adoptive parents. This study indicates a strong predilection for severe infection and septic shock based on the genetic background of individuals.

Over the last 10 yrs, a large number of candidate genes associated with increased risk of infection have been identified, and many more will undoubtedly be discovered by mining the human genome (^{18, 19}). Most of the genetic traits associated with severe infection are readily evident and relate to defects in innate immune responses such as complement deficiencies (²⁰), neutrophil defects (²¹), and alterations in signal transduction (^22-24), pattern recognition molecules (^25-27), and polymorphisms, resulting in low levels of mannose-binding lectin (^28-30). Intriguing additional associations have been found between genetic determinants of cytokine responsiveness and the risk of death from severe sepsis (^{22-24, 30-32}). Genetic variants in the coagulation pathways and fibrinolytic pathways and the risk of mortality from systemic infection have also been identified (³³). These relationships further support the evolutionary link between and clinical significance of the coagulation system and innate immune responses in sepsis (^{6, 34}).

A myriad of acquired defects in innate immunity and host defense mechanisms against microbial pathogens also predispose to systemic infection and severe sepsis (^{6, 7}). Severe trauma, burns, breaks in the integument and mucous membrane barriers, obstructive lesions, and a multitude of immunosuppressive disease processes may affect the risk of severe infection and septic shock (^{4, 5}). The degree to which these elements and potentially other predisposing factors alter the risk of infection and the systemic host response to invasive infection is subject for detailed clinical investigation in the future. It is a rare patient indeed who truly develops sepsis without some type of predisposing factor leading to invasive infection and systemic inflammation.

I: Infection

The site of infection responsible for the induction of severe sepsis has repeatedly been demonstrated to have an important effect on outcome. Patients with sepsis originating from pulmonary infections, gastrointestinal infections, and central nervous system infections experience a significantly higher mortality rate when compared with septic patients with infections originating from the genitourinary tract or skin and soft-tissue infections (^{2, 15, 34, 35}). Bacteremic patients generally do worse than nonbacteremic patients, although this relationship is not found with all organisms or in all studies of sepsis (^{34, 35}). Primary bacteremia (in particular, bacteremia arising from contaminated intravascular devices) has a more favorable prognosis than secondary bacteremias (bacteremias that occur from a demonstrable primary focus such as the lung or intraabdominal abscess). This relationship is principally related to the ease of removal of the source of primary bacteremia vs. secondary bacteremia. Moreover, the quantitative level of pathogens present in the blood and the tissues is generally higher in secondary bacteremia than primary bacteremia.

The quantity of microorganisms and their intrinsic virulence are both important determinants of outcome in presence of invasive microbial infection (^{36, 37}). In a recently completed extensive review of the microbiology of severe infection and septic shock, Cohen et al. (³⁸) compiled >55,000 infections reported in the literature during the past 30 yrs, encompassing 510 references. The microbiological review demonstrated marked differences in survival statistics based on the identity of the infecting microorganism and the site of infection (Fig. 1). More virulent organisms, such as Staphylococcus aureus, Streptococcus pyogenes, and Pseudomonas aeruginosa, have a higher mortality rate compared with less pathogenic organisms (i.e., coagulase-negative staphylococci, Acinetobacter species) over a wide range of sites of infection.

Figure 1

Clearly, the nature of the infecting organism that causes sepsis is not sufficient alone to determine outcome because the host response to infection is also critically important in determining the mortality rate from systemic infection. Candida species has a particularly poor prognosis when found in substantial concentrations in the blood stream or in major organs. The fact that candidal infections are observed principally in immunocompromised patients or chronically ill patients with nosocomial infection probably accounts for much of the excess mortality due to this fungal organism rather than the inherent virulence of Candida species per se (³⁴). A valuable feature of PIRO is that it is designed to account for both the type of infecting organism and the nature of the host response in the host-pathogen interaction in severe sepsis.

The relative contribution of the pathogen vs. the host in determining outcome in sepsis needs to be specifically analyzed. As an example, Candida sepsis from a contaminated vascular catheter in an otherwise healthy patient after a surgical procedure will likely have a considerably lower risk of death than a patient with refractory leukemia with disseminated candidiasis. What is the magnitude of this difference when compared with a patient with coagulase-negative staphylococci in the blood from a contaminated vascular catheter, or a leukemic patient with bacteremia from the same organism? It is very likely that the outcome of sepsis is predicated by both the type of infecting organism and the host's capacity to respond to it. The relative contributions of each element of the septic process need to be characterized using the PIRO format.

R: Response

The host inflammatory response to systemic infection is perhaps the key element to understanding severe sepsis and septic shock. The complexities of the host response and the variations between patients in response to microbial challenge present the major hurdle to a comprehensive understanding of the pathophysiology of septic shock (^{4-6, 35}). Although many of the essential cellular and humoral elements that contribute to septic shock are increasingly understood, integration of these disparate processes into a clearly defined and recognizable system has proven to be extremely difficult. It is recognized that multiple clinical and laboratory variables have the capacity to alter the host response to infection.

Many of these are rather simple and straight forward. The patient's age, nutritional status, sex, genetic background, and underlying disease processes may affect the host innate immune and acquired immune response to invasive infection (^39-43). The complexity arises from the redundancy of inflammatory networks and the fact that there are multiple, nonlinear, competing negative feedback loops and amplification pathways that occur simultaneously in the septic host. There is also evidence of compartmentalization at the immune response where cellular elements in the systemic circulation may express a different series of immune responses when compared with cells residing in the liver, spleen, lung, or other host tissues.

Clearly, many of the elements that encompass predisposing factors affect the host response to sepsis. Overlap between various components of PIRO is to be expected. In specific patients, one element may dominate over the others, whereas other patients may experience equal contributions from each element of PIRO to determine outcome. Deciphering the complexities of the host response to systemic microbial challenge is incomplete at best and remains a major focus on biomedical research in the field of sepsis at the present time (^{6, 44}).

It is recognized that some patients have a marked systemic response, often referred to as hyperinflammation, induced by certain microbial pathogens (^{6, 35}). The clearest examples of this situation are observed in previously healthy children who develop meningococcal sepsis (⁴⁵) or patients with streptococcal toxic shock syndrome after a relatively mild soft-tissue injury (^{37, 38}). The other extreme exists in febrile, neutropenic cancer patients or bone marrow transplant patients who have a markedly impaired systemic host response to microbial infection and often die of overwhelming infection with an inadequate host response (^{5, 6}). There are numerous gradations in the spectrum between these two extremes, yet it is often difficult to clinically distinguish between patients with hyperinflammation or inadequate inflammatory responses. Sepsis itself can induce a state of relative immune suppression. Late-onset, systemic infections from Candida species, coagulase-negative staphylococci, enterococci, Acinetobacter species, and other relatively avirulent organisms are well-recognized complications in patients after prolonged intensive care unit stays after a severe community-acquired infection.

It seems naive and potentially dangerous to lump patients together with hyperinflammation and immune suppression under the general title of septic shock and attempt to treat each of them with the same immunomodulatory treatment. Some patients may need immune reconstitution and immunoadjuvants, whereas others may benefit from anti-inflammatory strategies (^{6, 25, 30}). Improvements in rapid, functional genomics and proteomics may allow patients to be categorized based on their innate host immune and coagulopathic response to infection in the near future. A better understanding of the state of the host response to systemic infection will greatly facilitate the use of appropriately targeted therapies for septic patients in the future.

O: Organ Dysfunction

Organ dysfunction is the final tissue sequelae in response to severe sepsis and the ultimate determinant of survival. It has been amply demonstrated that septic hosts who have progressive multiple organ failure are much more likely to succumb to severe sepsis than those who develop a single or no organ dysfunction in response to sepsis (^{10, 13, 42-46}). What remains unclear is the reason or reasons why some patients develop one type of organ dysfunction and others develop another organ dysfunction, despite seemingly similar septic stimuli. Some patients develop profound coagulation abnormalities in the early phases of sepsis, whereas others develop severe acute respiratory dysfunction syndrome, whereas others experience acute renal failure (^{4, 5}). The sequence and pattern of organ dysfunction after severe sepsis may also vary between patients and provide some insights into the likelihood of survival and response to therapy.

Another important question in sepsis research is the effect of preexisting disease processes on the risk of organ dysfunction during septic shock. What is the likelihood that specific preexisting organ dysfunction will predispose these organs to fail in the presence of sepsis? It is often difficult to determine whether progressive organ dysfunction is related to underlying disease process or to the septic process. This is particularly problematic when primary organ infection is complicated by primary organ dysfunction from sepsis. Secondary organ injury remote from the site of the initial septic focus of infection is readily distinguishable from primary organ injury. The diffuse lung injury that occurs after localized pneumonia may be more difficult to reconcile than a patient with ascending cholangitis who develops secondary acute lung injury. It is clear that endothelial cells lining capillary beds from different organs express differential responses to septic stimuli. This heterogeneity within the microcirculation may account for some of the differences in organ dysfunction seen in patients with sepsis (⁴⁷). The loss of regulation and control of organ function in severe sepsis is certainly an area of active research in the understanding of sepsis at the present time.

Conclusion

PIRO is an innovative methodology by which to reexamine the pathophysiologic events that underlie septic shock (Table 1). Whether this system will ultimately prove to be more useful than general terms such as sepsis, severe sepsis, or septic shock remains to be seen. The TNM classification for understanding cancer biology has proven invaluable for some neoplasms and has been of limited or no benefit for others. It is likely that PIRO will also prove to be a useful method for analyzing certain aspects of sepsis but may not be of clinical utility in all aspects of sepsis' physiology. One of the favorable attributes of PIRO is that its elements are readily testable in clinical and basic research in sepsis. This methodology has yet to be fully put to the test as a novel strategy for analyzing sepsis.

Table 1

General sepsis studies are now giving way to clinical investigations of specific patient populations. Focused trials in severe community-acquired pneumonia, postoperative peritonitis, meningococcal sepsis, neutropenic sepsis, and sepsis in the neonate are examples of clinical investigations that acknowledge differential elements in determining outcome. The splitter approach may be preferable to the clumper approach to sepsis trials until the entire septic process can be better integrated. Current sepsis trials in the planning stages will analyze some components of PIRO, and it is hoped that this will serve as the impetus to fully develop this concept in future investigations in septic shock.